A hybrid algorithm to solve the flexible oil refinery production planning problem

The goal of production planning in a refinery is to produce as many valuable products as possible such as gasoline, jet fuel, diesel, etc., while meeting market demand and other constraints. Crude oil refining is one of the most complex chemical industries; Therefore, optimizing the production planning of an oil refinery is considered as one of the most difficult and challenging issues in this field. Due to rapid changes in industry-related technologies such as the construction of new catalysts, the design of more flexible process units, the flexibility of refineries is increasing rapidly. With the flexibility of refineries, their production planning requires a mathematical model that can be used to make the best decision at the right time to meet market demand at the lowest production cost. In this paper, the production planning of a flexible refinery is modeled using mathematical relationships between macro parameters such as product demand, production conditions, fixed and variable production costs, and inventory costs of petroleum products. To solve it, a hybrid algorithm of combining genetic algorithm and fix and optimize is proposed. The results with using of 63 simulated problems in small, medium and large dimensions show that the near-optimal solution obtained from the hybrid method deviates 0.23 and 0.12 percent of the exact solution of the problem on average in small and medium dimensions respectively. Also in large dimensions where it was not possible to calculate the exact answer by the computer, this algorithm can answer in an average of 87 seconds.